* Below is an edited excerpt from Dr. Michael Sudduth's forthcoming book, The Reformed Objection to Natural Theology. It is posted here with his permission
*An excerpt from Michael Sudduth, The Reformed Objection to Natural Theology (2006 draft)
Section III of chapter 13
Epistemic Certainty and Belief in God
Many theists maintain that they are certain of the truth of various theological propositions, among them being the proposition that God exists. I want to argue that for at least one important sense of certainty this position is false. The relevant sense of certainty here is what is called epistemic certainty, a species of certainty distinguished from so-called psychological certainty. The latter is merely descriptive and refers to a cognizer having maximal conviction or assurance of the truth of some proposition. While many theists are psychologically certain of God’s existence, this is epistemologically uninteresting. People have psychological certainty regarding all sorts of false propositions (e.g., Santa Claus exists, the world is flat, Elvis is alive). By contrast, a belief that is epistemically certain has some epistemic merit or credential, an epistemic merit or credential that is in some respect unsurpassed by other beliefs. I’ll argue that theistic belief (and belief in other theological propositions) is not epistemically certain.
I. Epistemic Certainty as Indubitability
Most accounts of epistemic certainty are tied to beliefs having some sort of epistemic immunity. This is typically articulated in terms of immunity from doubt or indubitability, where this means roughly the absence of any grounds for doubting a proposition p or doubting that one knows that p.1 Recall here that we’ve been working with a view of propositional knowledge according to which a person S knows that p just if (A) S firmly believes that p (belief condition), (B) S’s belief that p is (adequately) warranted (warrant condition), and (C) S’s belief that p is true (truth condition). More precisely, then, we can say that if a proposition p is epistemically certain for some person (in the sense of being epistemically indubitable), then it is logically impossible for (A) to be satisfied and for there to be any grounds for doubting that either (B) or (C) is satisfied. Since (B) and (C) are necessary conditions for knowledge, grounds for doubting either (B) or (C) will constitute grounds for doubting whether the persons knows that p.
A fairly crucial question, though, is what constitutes a ground or reason for doubting some proposition p? A baseline requirement is typically that there is no proposition q such that
(1) adding q to S’s beliefs would at least lower the warrant of S’s belief that p (ever so slightly).2
To this is frequently added:
(2) q is epistemically possible3 for S, meaning either that:
(a) S lacks conclusive reasons for believing the denial of q
(b) S is not warranted in believing the denial of q.
So a person S’s belief that p will be epistemically certain only if4 there is no epistemically possible proposition q for S that would result in lowering the warrant of S’s belief that p ever so slightly if q were added to S’s beliefs.
Is theistic belief indubitable and hence epistemically certain in this sense?
First, there are presumably many propositions that, if added to a person’s beliefs, would result in the warrant of theistic belief being lowered to some degree, however minimal: (i) theistic belief is produced by non-truth-aimed cognitive faculties (maybe in conjunction with the proposition “beliefs generated by non-truth directed faculties are unlikely to be true”), (ii) the concept of God involved in theistic belief is apparently inconsistent, (iii) there is Nth degree of apparently gratuitous evil in the world, (iv) the conjunction of theism and some other apparently true proposition is clearly logically inconsistent, (v) the conjunction of theism and some other true proposition is apparently logically inconsistent, and (vi) the sources of theistic belief among the various theistic religions and denominational divisions therein often yield incompatible propositions about God as output. Now we need not maintain that any of these propositions gives us good reasons to suppose that theism is false, or even good reasons for no longer believing that theism is true. They are simply some reason to suppose that theism is false or some reason for no longer believing that theism is true. Nor is it necessary that any of these propositions, if accepted, lower the warrant of theistic belief significantly, but only that they do so ever so slightly. Hence, if knowledge does not require epistemic certainty, then satisfying conditions (1) and (2) would be consistent with knowing theism to be true. The point here is not that theistic belief has no significant epistemic credentials, but only that it does not have the best possible epistemic credentials. All that is necessary, then, is to find propositions that would suffice to lower the warrant of theistic belief ever so slightly, and these would be any proposition that at least minimally counts against the truth of theism or at least minimally counts against our thinking that theism is true.
Secondly, since doubt makers need not have significant epistemic credentials, it seems fairly clear that there is going to be some proposition q that both lowers the warrant of theistic belief if added to S’s belief and is epistemically possible for S, at least for most cognizers. Indeed, the candidate doubt maker need not be warranted at all. It must only be the case that we are not warranted (conclusively or otherwise) in believing their denials. But it certainty seems that at least some of the relevant propositions above are such that we lack conclusive reasons for believing their denials. Indeed, we might even suppose that (iii), (iv), (v), and (vi) have some degree of warrant for most cognizers.
But isn’t it the case that if a person were warranted in her theistic belief, then she would be warranted in denying any proposition logically incompatible with theism? And in that case, condition (2) will not be satisfied. Apart from the fact that this would be applicable only to epistemic certainty cashed out in terms of the conjunction of (1) and (2b), it is misguided for at least two other reasons. First, since doubt makers need not be reasons for believing that theism is false, many of them will be compatible with the truth of theism and even compatible with S’s theistic belief being warranted. So even if a person’s being warranted in theistic belief entailed that she is warranted in denying every proposition logically incompatible with theism, she would not on that account be warranted in denying every candidate doubt maker for theism, as in the case of (iii), (iv), (v), and (vi) above. Secondly, the supposition that if a person is warranted in believing that p, then the person is warranted in affirming every entailment of p (and thus warranted in denying every proposition that is logically incompatible with p) is questionable.5 This so-called “entailment principle” overlooks the possibility that my warrant for believing p may be less than my warrant for believing some proposition that is incompatible with p. In that case, while I may indeed be warranted in believing p, I’m not necessarily warranted in disbelieving every proposition that is incompatible with p. I might of course if theistic belief had maximal warrant. But is this so?
II. Epistemic Certainty as Maximal Warrant
The idea of “maximal warrant” introduces a more modest way to think of epistemic certainty. On this view, a proposition p is epistemically certain for a person S just if S is warranted in believing p and there is no other proposition q that has more warrant for S than his belief that p.6 Now an indubitable belief will be maximally warranted, but a belief can be maximally warranted even if it is not indubitable. After all, there might be no proposition q more warranted or better justified than p for a person S, but it still may be that q is epistemically possible for S and such that q would lower the warrant of S’s belief that p if q were added to the rest of S’s beliefs. Indeed, there might be some other proposition q, which has as much warrant for S as S’s belief that p, but where q is doubt-maker for p. So this is a more modest account of certainty. It does not entail indubitability. We might suppose, then, that while theistic belief is not indubitably certain, there is no belief that has more warrant for a person than theistic belief, and so theistic belief is certain in this more modest sense.
One of the problems with this proposal, though, is that it appears that there are beliefs that have more warrant than any paradigm case of theistic belief. For example, take any number of introspectively evident beliefs about one’s current states of consciousness, e.g., I feel tired, it seems to me that there is a computer in front of me, I am thinking now. It is generally held that these sorts of beliefs have fairly strong epistemic credentials, perhaps the strongest sort human beliefs can have. Different reasons have been proposed for supposing why this is the case, most of which involve different accounts of how introspective beliefs involve particular entailments between the conditions of knowledge. One might suppose, for example, that introspective beliefs can’t be true without our believing them and we can’t believe them without their being true. So the truth condition entails the belief condition, and conversely. One might further suppose that in this case introspective beliefs enjoy the best sort of epistemic credentials possible. After all, warrant is a truth-indicating property. Perhaps there is no better indication that a proposition is true than if its truth is entailed by our simply believing it. Alternatively, we might suppose that introspective beliefs that are in fact based on self-presenting states of consciousness are based on grounds that guarantee the truth of the belief, for in this case the ground of the belief would be what makes the belief true. Hence, for introspective beliefs the warrant condition (being connected to the grounds of a belief) entails the truth condition, and this might be viewed as the best conceivable sort of warrant. Finally, on other accounts introspective beliefs are warranted simply by our holding them, even if their truth is not entailed either by their being held or warranted.7 In each of these three cases, we see ways of explaining why introspective beliefs enjoy a sort of privileged epistemic status.
Now it is fairly obvious that theistic beliefs are not beliefs about our current states of consciousness, but neither they do appear to have the sort of epistemic credentials that such beliefs possess. Consider the three accounts of these credentials. First, unless one adopts a radically anti-realist view of God, theistic beliefs are not made true by our believing them, nor does the truth of any theistic proposition guarantee that anyone will believe it.8 Secondly, theistic beliefs are not based on grounds that make theistic belief true and thus that guarantee the truth of beliefs based on such grounds. The putative grounds of theistic belief, be it religious experience, intuition, the sensus divinitatis, inference from various features of the world, are not identical to the fact that makes theism true. Are theistic beliefs the sort of beliefs that would at least be warranted simply by virtue of our holding them? This seems doubtful, even in those cases where the beliefs are true theistic beliefs. Perhaps a person comes to hold some true theistic belief on the grounds that he has communicated with apparitions from beyond the grave who have spoken to him of the beauties of heaven, but he is simply suffering from a mental disorder. From the vantage point of most externalist and internalist epistemologies, it’s hard to see how the person’s belief would be warranted in this circumstance.9 The warrant of theistic beliefs is not just given by the mere fact that one holds such beliefs. It is thus hard to see how theistic beliefs can be warranted to the same degree as beliefs about our current states of consciousness.10
One might suppose, though, that a different answer can be drawn from Plantinga’s epistemology. On Plantinga’s view, a person whose relevant cognitive faculties are functioning properly will hold a firm theistic belief that has a high degree of warrant. In fact, on Plantinga’s view, theistic belief is indefeasible for all fully rational persons. No proposition a fully rational person entertains could serve as a defeater for theistic belief. That’s a pretty substantial epistemic credential. Of course, defeaters against theistic belief exist according to Plantinga, but only because the epistemic integrity of some other aspect of our cognitive establishment (perhaps the sensus divinitatis) has been compromised, say by the noetic effects of sin. It may very well be true that apart from the noetic effects of sin, humans would believe in God just as firmly as they believe in their own existence, the existence of an external world, other minds, and various a priori truths, and perhaps our theistic beliefs would be just as warranted as these other beliefs. But this is an ideal view of the human cognitive situation, at best true for some original cognitive design plan and perhaps true for us in our final state. But now we see through a glass darkly, as it were. As indicated in prior chapters, the noetic effects of sin are a factor in assessing the degree to which all our beliefs can be warranted, including belief in God. It is hard to see how theistic belief can be maximally warranted for humans under any post-lapsarian cognitive design plan.11
So I think we must conclude that there isn’t a very strong case for supposing that theistic beliefs are epistemically certain in either the sense of indubitability or maximal warrant. In fact, this looks just plain false.
III. The Senses in which Belief in God is Certain
In what sense, then, can theistic belief be certain?
Many theists are psychologically certain of the existence of God and other theological propositions. However practically useful such a belief is, psychological certainty says nothing about the normative axis of belief, the epistemic merits or credentials of a belief. So we must look elsewhere for a relevant and plausible sense in which theists may have certainty concerning the existence of God and other theological propositions.
If God’s existence is logically necessary, then theistic belief is certain in a purely logical sense, for then it will not be logically possible to believe that God exists and for this belief to be false.12 But this isn’t epistemic certainty. Since it is logically possible to believe a logically necessary truth and yet not know the proposition, or even be warranted in holding it, clearly there is a sense in which it is impossible to be mistaken in a belief and yet for this to carry no epistemic significance. Suppose Jack believes nothing is red and non-colored because a character in a cartoon asserts it and Jack is inclined to accept whatever he hears cartoon characters affirm. His belief is true, but it would seem to have little by way of warrant. The logical status of the proposition tells us nothing about the positive epistemic status of his belief in the proposition.13
I would suggest that the relevant and plausible kind of certainty is moral certainty. A morally certain belief is beyond all reasonable doubt, though not beyond all possible doubt. In positive terms, such beliefs are highly probable. Morally certain beliefs entitle us to be sure about our beliefs, and at least some of them they carry a degree of warrant that is plausibly sufficient, together with the satisfaction of the truth condition, for knowledge. Thus morally certain theistic beliefs do justice to the Biblical passages that suggest Christians ought to be sure about their faith and that Christians have knowledge of God.14
1 This is not the same as saying that it psychologically impossible to doubt the proposition. Clearly enough it might be psychologically impossible to doubt a proposition even if there are grounds for doubting it. Many children find it impossible not to believe in Santa Claus. Grounds or reasons for doubting his existence are not hard to come by. Indubitability is a normative concept.
2 This need not be a reason for supposing that the denial of p is true. It can simply be a reason for our no longer thinking that p is true.
3 (2) shows us that a doubt raiser need not have any significant epistemic credentials. It is sufficient that a statement’s negation not have any significant epistemic credentials. Recall that Descartes conceded that doubt makers might be very doubtful themselves, but this is acceptable when searching – as Descartes was - for the least or slightest ground of doubt. See E.M. Curley, Descartes against the Skeptics (Cambridge, MA: Harvard University Press, 1978), pp. 85-86, and Jeffrey Tlumak, “Certainty and Cartesian Method,” in Descartes Critical and Interpretive Essays, ed. Michael Hooker (Baltimore: Johns Hopkins University Press, 1978), pp. 44-53.
4 Whether such conditions are sufficient is another issue. Klein, for example, thinks that the idea of absolute certainty requires a third condition to the effect that there be no true proposition, d, such that if D were added to S’s beliefs, the warrant for believing p would be reduced. With this addition, certainty will entail truth. See Klein, “Certainty,” in A Companion to Epistemology ed., Jonathan Dancy and Ernest Sosa (Oxford: Blackwell Publishers, 1992), pp. 61-64.
5 To cite a counter-example proposed by Audi (Belief, Justification, and Knowledge (Belmont, CA: Wadsworth Publishing Company, 1988), p. 78). Suppose I add up some figures in column and believe that the total is 10,395. Suppose that I am correct and this belief is warranted. Suppose that I know I am correct. It follows, though, that the total is 10,395, even if I made a mistake in my addition. But it seems implausible to suppose that I could be warranted in believing this consequence, much less know it. Hence, we are not necessarily warranted in believing every entailment of propositions we are warranted in believing.
6 For instance, according to Roderick Chisholm: “p is certain for S = Df For every q, believing p is more justified for S than withholding q, and believing p is at least as justified for S as is believing q.” (Chisholm, Theory of Knowledge, 3rd edition (New Jersey: Prentice Hall, 1989), p. 12). See also Feldman, “Epistemic Appraisal and the Cartesian Circle,” Philosophical Studies 27 (1975), p. 43.
7 See Alston, “Self-Warrant” in Alston, Epistemic Justification (Ithaca: Cornell University Press, 1989). Cf. Alston, “Concepts of Epistemic Justification,” p. 106 in same volume for the contrasting theory of truth-warrant.
8 My believing that there is a God makes it true that “Michael Sudduth believes that there is a God”. But the proposition that is made true by my act of believing that there is a God is not a proposition about God. It is a proposition about Michael Sudduth’s mental state.
9 The actual grounds of the belief in this situation would not satisfy either a proper function or reliability constraint on warrant. So the externalist must reject the idea that theistic belief is warranted in this situation. Of course, an internalist might concede that in this situation a person has evidence for the truth of his belief and thus is justified or rational in some weak sense. But is this enough for warrant? No. There is some true proposition, such that if the person were to believe it, he would no longer be justified (in an internalist sense) in his belief, namely that proposition describing the fact that he suffers from a cognitive disorder. Hence, from an internalist perspective, the person’s belief would not pass the test of an indefeasibility clause designed to rule out Gettier cases. So there is a clear sense in which the person’s belief would not be warranted even from an internalist perspective, whatever other epistemic merits it might have.
10 Something similar should probably be said for at least some self-evident truths known a priori, for instance those who negations are evidently self-contradictory and so must be false. Even if the denial of theism contains an implicit contradiction, the contradiction isn’t as evident as the denial of, say, “all red things are colored.”
11 Perhaps there are exceptions here, e.g., the beliefs of Moses on Mt. Sinai, St. Paul on the road to Damascus, or the disciples with the resurrected Christ. But most people are not in this sort of privileged position.
12 This does not require that the denial of theism be self-contradictory. There are some logically necessary propositions whose negations are not self-contradictory, e.g., 2 + 2 = 4.
13 Of course if a person knows that God exists, then it is logically impossible for the person to be mistaken in this belief since knowing p entails (by definition) that S believes p and p is true. But in this sense all knowledge is infallible. There’s nothing unique here about the knowledge of God. Nevertheless, it is logically possible for a person to believe p and not know p. A more rigorous understanding of infallibility would be that a person’s belief that p entails that S knows that p. Where knowledge entails that the person is warranted in the belief and the belief is true, it would follow that an infallible belief is true and warranted. But as far as I can see theistic belief is not infallible in this sense, for reasons already noted above in text.
14 Turretin distinguished between three kinds of certainty: mathematical, moral, and theological. Institutio theologiae elencticae, II.iv.22. Turretin denied that theological propositions have mathematical certainty, and this would seem to conform to the negative axis of my argument. He also denied that the certainty of theological propositions is merely moral, and this would seem to conflict with the positive axis of my argument. Although theological certainty appears to be located in between mathematical and moral certainty, it isn’t adequately clear whether or how it is epistemically distinct from them. Turretin unpacks moral certainty as equivalent to conjecture, the acceptance of probabilities on the grounds of evidence. This may be a weaker view of moral certainty than the one I’m employing above. Moreover, if—as seems plausible—Turretin saw theological certainty as partaking of the psychological qualities of mathematical certainty and the epistemic qualities of moral certainty, then our accounts will be roughly identical.